Happened in AI

AI laws tracker

AI laws, regulations, and executive orders around the world: whether each is binding, where it stands, and the dates that matter.

Download: CSV · JSON · Last checked 22 Oct 2026

Next dates

  1. European Union: Artificial Intelligence Act (Regulation (EU) 2024/1689). End of the grace period for Article 50(2) machine-readable marking, for generative systems placed on the market before 2 August 2026 (added by the AI Omnibus).
  2. California, United States: Transparency in Frontier Artificial Intelligence Act (SB 53). Some provisions take effect; Office of Emergency Services begins publishing anonymized annual summaries of reported incidents.
  3. Colorado, United States: Colorado AI Act (SB 24-205), repealed and replaced by SB 26-189. SB 26-189 takes effect for consequential decisions made on or after this date.
  4. New York, United States: Responsible AI Safety and Education (RAISE) Act. Takes effect.
  5. Vietnam: Law on Artificial Intelligence (No. 134/2025/QH15). Transition period ends for most sectors.
  6. California, United States: Companion chatbots law (SB 243). Annual reporting begins.
  7. California, United States: Workplace AI package, including the No Robo Bosses Act (SB 947), AB 1883, and SB 951. SB 947 takes effect.
  8. Utah, United States: Artificial Intelligence Policy Act (SB 149, 2024), as amended by SB 226 and SB 332 (2025). Scheduled repeal date after the SB 332 extension.
WhereLaw or ruleStatusKey datesSources
Brazil AI bill (PL 2338/2023)Bill · Voluntary or non-bindingA risk-based framework for AI development and use in Brazil. Copyright treatment of training data is a major open issue.Still in a Chamber of Deputies special committee as of October 2026; several planned vote dates have passed. If the Chamber changes the text, it returns to the Senate. A government bill, PL 6237/2025, was attached in March 2026. ProposedChecked
  • Senate approves the substitute text.
  • Bill formally sent to the Chamber of Deputies.
  • Rapporteur Aguinaldo Ribeiro says the Chamber vote should wait until after the October 2026 elections.
 1 2
California, United States Companion chatbots law (SB 243)Law · BindingSets safeguards for operators of AI companion chatbots, including suicide-prevention protocols and annual reporting to the state Office of Suicide Prevention. Customer-service and other utility bots are exempt. Includes a private right of action. In forceChecked
  • Adopted
  • Signed by Governor Newsom.
  • Most deployment requirements take effect.
  • Annual reporting begins.
 1 2
California, United States Transparency in Frontier Artificial Intelligence Act (SB 53)Law · BindingRequires frontier AI developers to publish safety frameworks and report critical safety incidents to the state, protects whistleblowers, and sets extra duties for large frontier developers with over $500 million in annual revenue. Enforced by the attorney general through civil penalties.Story In forceChecked
  • Adopted
  • Signed by Governor Newsom.
  • Most provisions take effect.
  • Some provisions take effect; Office of Emergency Services begins publishing anonymized annual summaries of reported incidents.
 1 2
California, United States Workplace AI package, including the No Robo Bosses Act (SB 947), AB 1883, and SB 951Law · BindingSB 947 bars employers from relying solely on an automated decision system to fire or discipline workers and requires human corroboration and written notice. AB 1883 limits AI tools that infer workers' emotions or collect neural data. SB 951 requires layoff notices to disclose when cuts are caused by AI or automation.Some early coverage described a 1 January start for SB 947; the bill text sets 1 July 2027. Effective dates for AB 1883 and SB 951 were not confirmed in sources checked for this record; see the related story. Newsom vetoed an earlier version, SB 7, in October 2025.Story AdoptedChecked
  • Adopted
  • Signed by Governor Newsom.
  • SB 947 takes effect.
 1 2 3
Canada Artificial Intelligence and Data Act (part of Bill C-27)Bill · Voluntary or non-bindingWould have regulated high-impact AI systems at the federal level. Canada has no federal AI-specific law; AI use is governed by privacy laws such as PIPEDA and Quebec's Law 25 and by sector rules.Status is recorded as rejected because the bill lapsed at prorogation; it was not voted down. AI minister Evan Solomon has said AIDA will not be revived in its old form. RejectedChecked
  • Bill C-27 dies on the Order Paper when Parliament is prorogued.
  • Government tables Bill C-36 (privacy and consumer data), which does not contain a general AI framework.
 1 2 3
China Interim Measures for the Management of Generative Artificial Intelligence ServicesRegulation · BindingApplies to generative AI services offered to the public in China, setting duties on content, training data, and labeling, and requiring providers to explain training data and algorithms to regulators on request. Internal research and enterprise use not offered to the public are excluded.Issued by the Cyberspace Administration of China with six other agencies. Sources disagree on the issue date (10 July or 13 July 2023), so the adopted date is omitted. In forceChecked
  • Takes effect.
 1 2
China Measures for Labeling AI-Generated and Synthetic Content, with mandatory standard GB 45438-2025Regulation · BindingRequires AI service providers to add visible labels to AI-generated text, images, audio, video, and virtual scenes, and embedded labels in file metadata; app platforms must ask developers whether they offer AI generation services.Jointly issued by the CAC, MIIT, Ministry of Public Security, and NRTA. Sources give different issue dates in March 2025 (7, 14, or 17 March), so the adopted date is omitted. English translations of the title vary. In forceChecked
  • Measures and the companion mandatory national standard take effect.
 1 2
Colorado, United States Colorado AI Act (SB 24-205), repealed and replaced by SB 26-189Law · BindingCovers automated decision-making technology used in consequential decisions such as employment, housing, and lending. SB 26-189 drops the original duty of care, impact assessments, and risk-management programs, and instead requires disclosures plus consumer rights to access, correction, and human review of adverse outcomes. The attorney general enforces; there is no private right of action.The effective date changed twice: 1 February 2026 (original), then 30 June 2026 (SB 25B-004), then 1 January 2027 (SB 26-189). The court's stay reportedly extends to successor legislation, xAI has signaled a challenge to SB 26-189, and one firm reports the attorney general will not enforce until rulemaking is complete, so the 2027 date could slip. One source gives 1 May 2026 for the SB 26-189 signing; most give 14 May. DelayedChecked
  • Adopted
  • SB 24-205 signed, with an effective date of 1 February 2026.
  • SB 25B-004 signed after a special session, moving the effective date to 30 June 2026.
  • xAI files a federal constitutional challenge.
  • Federal court stays enforcement pending a preliminary injunction ruling.
  • SB 26-189 signed, repealing and replacing SB 24-205 with a narrower disclosure and consumer-rights law.
  • SB 26-189 takes effect for consequential decisions made on or after this date.
 1 2 3
Council of Europe (international) Framework Convention on Artificial Intelligence and Human Rights, Democracy and the Rule of Law (CETS No. 225)Treaty · BindingThe first legally binding international treaty on AI, requiring parties to keep AI activities consistent with human rights, democracy, and the rule of law.Not in force as of this check. It enters into force on the first day of the month after three months following five ratifications, including at least three Council of Europe member states. Trackers reviewed in September and October 2026 recorded only the EU's ratification; one 2025 secondary claim of entry into force on 1 November 2025 was not supported. Check the Council of Europe Treaty Office for the current count. AdoptedChecked
  • Opened for signature in Vilnius; first signatories included the UK, the US, the EU, Israel, Norway, and others.
  • European Union deposited its consent to be bound (reported by secondary trackers).
 1 2
European Union Artificial Intelligence Act (Regulation (EU) 2024/1689)Regulation · BindingSorts AI uses by risk, bans some practices, sets requirements for high-risk systems, and places duties on providers of general-purpose AI models placed on the EU market.The adopted date is the date the regulation was signed. Regulation (EU) 2026/1744 (the AI Omnibus), in force since 27 July 2026, postponed the high-risk dates: Annex III from 2 August 2026 to 2 December 2027, and Annex I from 2 August 2027 to 2 August 2028. The obligations were postponed, not removed. Article 50 transparency duties were not deferred, apart from the marking grace period noted above.Story Partly in forceChecked
  • Adopted
  • Published in the Official Journal of the European Union.
  • Entered into force.
  • Prohibited practices and general provisions apply.
  • Obligations for providers of general-purpose AI models apply.
  • Most remaining provisions apply, including Article 50 transparency duties; the Commission and AI Office can enforce general-purpose AI obligations, including fines.
  • End of the grace period for Article 50(2) machine-readable marking, for generative systems placed on the market before 2 August 2026 (added by the AI Omnibus).
  • General-purpose AI models placed on the market before 2 August 2025 must comply.
  • High-risk obligations for stand-alone (Annex III) systems apply. Originally 2 August 2026; moved by the AI Omnibus.
  • High-risk obligations for AI in regulated products (Annex I), such as medical devices and machinery, apply. Originally 2 August 2027; moved by the AI Omnibus.
 1 2 3 4
European Union Digital Omnibus on AI (Regulation (EU) 2026/1744), amending the AI ActRegulation · BindingAmends the AI Act: postpones high-risk application dates, adds prohibitions on AI-generated non-consensual intimate imagery and child sexual abuse material, revises the AI literacy obligation, adds simplifications for smaller companies, and strengthens the AI Office's supervisory powers.The adopted date is the date of signature shown on EUR-Lex. Some secondary sources also describe the regulation as applying from 2 August 2026; entry into force was 27 July 2026.Story In forceChecked
  • Adopted
  • European Commission proposal.
  • Provisional political agreement between Parliament and Council.
  • Published in the Official Journal.
  • Entered into force
 1 2 3
European Union General-Purpose AI Code of PracticeGuidance · Voluntary or non-bindingA voluntary code, written by independent experts, that general-purpose AI providers can sign to demonstrate compliance with the AI Act's transparency, copyright, and safety obligations. Signatories include Amazon, Anthropic, Google, IBM, Microsoft, Mistral AI, and OpenAI; xAI signed only the safety and security chapter.Signing is voluntary, but the AI Act obligations it maps to are binding. Non-signatories must show compliance by other adequate means. The signatory list changes over time.Story AdoptedChecked
  • Adopted
  • Final code published by the AI Office.
  • Commission opinion reportedly found the code adequately covers AI Act Articles 53 and 55.
  • AI Act obligations for general-purpose AI providers begin to apply.
 1 2
Illinois, United States Human Rights Act amendment on AI in employment (HB 3773)Law · BindingBars employers from using AI in recruiting, hiring, promotion, discipline, or discharge in a way that discriminates against protected classes, bans using zip codes as a proxy, and requires notice to workers when AI is used.Sources reviewed did not confirm whether the Illinois Department of Human Rights has finalized implementing rules.Story In forceChecked
  • Adopted
  • Signed by Governor Pritzker.
  • Takes effect.
 1 2
India Information Technology (Intermediary Guidelines and Digital Media Ethics Code) Amendment Rules, 2026Regulation · BindingRequires online intermediaries to label synthetically generated audio, visual, and audio-visual content made to look real, attach provenance metadata where feasible, and act faster on takedowns. Text-only AI output is outside the definition.India has no AI-specific statute. Sources differ on the takedown window for the most harmful content, so it is not recorded here. The final rules dropped a draft requirement that labels cover 10 percent of the screen. In forceChecked
  • Adopted
  • Notified by the Ministry of Electronics and Information Technology.
  • Takes effect.
 1 2
Italy Law No. 132/2025 on artificial intelligenceLaw · BindingThe first national AI law in an EU member state. It complements the EU AI Act with sector rules, designated authorities, copyright and criminal provisions, parental consent for AI use by children under 14, and delegates much of the detail to implementing decrees.Sources differ on the Gazzetta issue number (223 or 224). Some implementing decrees are not yet in force. In forceChecked
  • Adopted
  • Final approval by Parliament.
  • Published in the Gazzetta Ufficiale.
  • Takes effect.
 1 2
Japan Act on Promotion of Research, Development and Utilization of AI-Related Technologies (AI Promotion Act)Law · BindingA promotion-focused framework that sets principles, a national AI plan, and roles for government and companies. It has no monetary penalties; the government can investigate rights-infringing uses of AI and respond with guidance and advice.One source reports that the Cabinet adopted the first AI Basic Plan on 23 December 2025; this was not confirmed by a second source and is not listed as a milestone.Story In forceChecked
  • Adopted
  • Passed by the Diet.
  • Most provisions take effect.
  • Remaining provisions take effect
 1 2
New York, United States Responsible AI Safety and Education (RAISE) ActLaw · BindingRequires large frontier AI developers to publish safety protocols and report safety incidents to the state within 72 hours, and creates an oversight office in the Department of Financial Services.The chapter amendment supersedes the December 2025 text. It replaced compute-cost thresholds with a $500 million revenue threshold and lowered maximum penalties from $10 million and $30 million to $1 million for a first violation and $3 million for later ones. AdoptedChecked
  • Adopted
  • Signed by Governor Hochul on condition of chapter amendments in 2026.
  • Chapter amendment signed (Chapter 96 of the Laws of 2026).
  • Takes effect.
 1 2 3
South Korea Framework Act on the Development of Artificial Intelligence and Establishment of Trust (AI Basic Act)Law · BindingSets a national AI governance framework, requires notice to users of generative and "high-impact" AI, imposes duties on high-impact AI providers, and requires qualifying foreign providers to designate a local representative. Maximum administrative fine is 30 million won.The adopted date is the promulgation date. Sources differ on the National Assembly passage date (26 or 27 December 2024). MSIT has said it will apply a grace period of at least one year before imposing fines, except in serious cases; the duties themselves apply from 22 January 2026. In forceChecked
  • Adopted
  • Promulgated as Act No. 20676.
  • President signs the Enforcement Decree (Presidential Decree No. 36053).
  • Act and Enforcement Decree take effect; MSIT issues guidelines on safety
 1 2 3
Texas, United States Texas Responsible Artificial Intelligence Governance Act (HB 149)Law · BindingApplies to those who develop, deploy, or market AI systems in Texas, with most duties aimed at government agencies and certain consumer-facing uses, and fewer burdens on private-sector development. In forceChecked
  • Adopted
  • Signed by Governor Abbott.
  • Takes effect.
 1 2
United Kingdom Pro-innovation approach to AI regulation (government response to the AI white paper)Guidance · Voluntary or non-bindingExisting sector regulators apply five cross-sector principles (safety, transparency, fairness, accountability or governance, and contestability) to AI within their own remits, supported by a central coordination function.The UK has no AI-specific statute. Labour's 2024 manifesto pledged binding rules for the most powerful models, but the May 2026 King's Speech did not include an AI bill, and reported earlier plans for frontier-model legislation have been shelved. The white paper date is recorded as the first of the month because the sources used here give only the month.Story AdoptedChecked
  • Adopted
  • AI regulation white paper and consultation published (month only).
  • Government response confirms a regulator-led approach with no new AI-specific law.
  • King's Speech includes no AI bill; it announces a Regulating for Growth Bill with powers for regulatory sandboxes, including for AI.
 1 2 3
United States (federal) Executive Order 14179: Removing Barriers to American Leadership in Artificial IntelligenceExecutive order · BindingDirects federal agencies to sustain US AI leadership, orders an AI action plan within 180 days, and requires review of policies issued under the revoked 2023 Biden order. In forceChecked
  • Adopted
  • Signed; directs review of actions taken under revoked Executive Order 14110 (October 2023).
  • Published in the Federal Register (90 FR 8741).
 1 2
United States (federal) Executive Order 14319: Preventing Woke AI in the Federal GovernmentExecutive order · BindingLimits federal agency purchases of large language models to those meeting two "Unbiased AI Principles" (truth-seeking and ideological neutrality), and requires contract terms to enforce them.Story In forceChecked
  • Adopted
  • Signed.
  • OMB issues implementation guidance on federal purchases of large language models.
 1 2
United States (federal) Executive Order 14365: Ensuring a National Policy Framework for Artificial IntelligenceExecutive order · BindingDirects federal agencies to challenge state AI laws seen as burdensome, to tie some federal funding to states' AI rules, and to propose federal preemption legislation. It does not itself invalidate any state law, and it exempts state laws on child safety, AI infrastructure, and state government procurement.Congress has not passed preemption. The Senate removed a ten-year state AI moratorium from the 2025 reconciliation bill by a 99-1 vote on 1 July 2025 (Senate record, also reported by Axios), and a preemption attempt was dropped from the FY2026 defense bill. Sources reviewed did not confirm whether the Commerce review or the related FTC policy statement has since been published.Story In forceChecked
  • Adopted
  • Signed.
  • Justice Department memorandum formally establishes the AI Litigation Task Force.
  • Deadline for the Commerce Department's review identifying "onerous" state AI laws; a March 2026 analysis reported it had not been released publicly.
  • Justice Department moves to intervene in xAI's challenge to the Colorado AI Act.
 1 2 3 4 5
United States (federal) Executive order on Promoting Advanced Artificial Intelligence Innovation and SecurityExecutive order · BindingAsks AI developers to let the government assess advanced models up to 30 days before wider release, on a voluntary basis, and directs agencies to prioritize AI-enabled cyber defense and enforcement against criminal uses of AI.The order binds federal agencies; developer participation in early access and benchmarking is voluntary, and the order states it does not authorize mandatory licensing or pre-clearance. A May 2026 draft reportedly sought 90 days of pre-release access.Story In forceChecked
  • Adopted
  • Signed.
 1 2
United States (federal) National Policy Framework for Artificial Intelligence (White House legislative recommendations)Guidance · Voluntary or non-bindingRecommends one national standard for AI that would preempt burdensome state laws while leaving state child-protection, fraud, consumer-protection, and zoning powers intact, and recommends against creating a new federal AI regulator.The adopted date is the release date. No federal preemption statute has passed as of this check. A September 2026 status report said the Obernolte-Trahan draft had stalled; sources conflict on whether a successor, the FRONTIER Act, was formally introduced in July 2026, so it is not listed as a milestone. ProposedChecked
  • Adopted
  • White House releases nonbinding legislative recommendations asking Congress to preempt state AI laws that impose undue burdens.
  • Representatives Jay Obernolte and Lori Trahan release a bipartisan discussion draft (Great American AI Act) that would preempt state AI development laws for three years.
 1 2 3
United States (federal) TAKE IT DOWN ActLaw · BindingCriminalizes publishing nonconsensual intimate images, including AI-generated "digital forgeries," and requires covered platforms to remove reported images and known identical copies within 48 hours of a valid request. In forceChecked
  • Adopted
  • Signed into law; criminal provisions on publishing nonconsensual intimate images take effect.
  • Platform notice-and-removal duties apply and FTC enforcement begins.
 1 2 3
Utah, United States Artificial Intelligence Policy Act (SB 149, 2024), as amended by SB 226 and SB 332 (2025)Law · BindingRequires businesses using generative AI to disclose it when a consumer directly asks, and to disclose proactively in high-risk interactions involving sensitive personal information or significant decisions, such as financial, legal, or medical advice. A separate 2025 law, HB 452, covers mental health chatbots.The act was originally set to repeal automatically in May 2025; sources disagree on the day (1 May or 7 May). SB 332 extended it to 1 July 2027, and SB 226 narrowed when disclosure is required. Signing and effective dates for the 2024 act were not confirmed in sources checked for this record, so they are omitted. In forceChecked
  • Scheduled repeal date after the SB 332 extension.
 1 2
Vietnam Law on Artificial Intelligence (No. 134/2025/QH15)Law · BindingSorts AI systems into three risk tiers and requires providers of medium- and high-risk systems to notify the Ministry of Science and Technology before deployment. Maximum fines are 2 billion dong for organizations and 1 billion dong for individuals.The list of high-risk systems and detailed classification criteria depend on implementing decrees that were still pending in sources checked. In forceChecked
  • Adopted
  • Passed by the National Assembly.
  • Takes effect.
  • Transition period ends for most sectors.
  • Transition period ends for healthcare
 1 2

Every row links to its sources (numbered links). Figures supplied by a company are marked as company figures. See How we report. Spotted an error? Reply to any newsletter issue.